A feasibility study on monitoring total phenolic content in sparkling wine press juice fractions using a new in-line system and predictive models

Ultraviolet and visible (UV–Visible) spectroscopy in conjunction with chemometrics modelling, was successfully applied to objectively differentiate grape juice press fractions (Vitis vinifera, cultivar Pinot Noir) for sparkling wine production. Two measurements modes were applied: (i) reflectance using a fiber optic probe in-line and (ii) transmission using a benchtop spectrophotometer. Different wavelength ranges for UV–Visible spectroscopy were evaluated and their ability to measure total phenolic concentrations in grape juice press fractions was compared. The differentiation of free run, early and late press fractions shows promise as a tool for the rapid discrimination of juice fractions when grapes for sparkling wine are pressed. Calibrations for total phenolics were prepared from press fraction spectral data using partial least squares (PLS) regression with a large number of wavelengths (230–700 nm) and multiple linear regression (MLR) using a small number of key wavelengths (230, 240, 280, 290, 520 nm). Calibration performance for both reflectance and transmission spectra was similar, but the best performing calibration used reflectance spectra at 240 and 290 nm (R2val = 0.95; SECV = 0.023 g/L; CV = 4.2%). Reflectance spectroscopy can thus be used in-line to predict total phenolics in grape juice with an acceptable accuracy and to discriminate press fractions. Insights arising from this research suggest a future possibility of objective, real-time discrimination of grape juice for better process control, monitoring and optimization of winemaking practices.
Keywords: UV–Visible spectroscopy; Total phenolic monitoring; In-line monitoring; PLS regression; MLR calibration; Optical sensor; Pinot noir